Atmospheric mesoscale (100's of meters to a few kilometers) temperature structure and the structure associated with thin cirrus and aerosol layers in the upper stratosphere and lower mesosphere are difficult to measure by ground and satellite based techniques. We show in this paper that the altitude range between about 10 and 80 km is amenable to satellite sounding techniques in the UV-visible-near infrared bands (approximately 200 to 900 nm). The rapid change in optical depth vs. line-of-sight (LOS) end point along a downward-viewing LOS in the 200 - 350 nm spectral range allows separation of atmospheric regions according to the LOS optical weighting functions. The UV imager weighting functions (200 - 300 nm) in combination with the satellite- sensor zenith angle effect allows sounding in the approximately 40 to 80 km region, while the visible band imagery allows detection and separation of high altitude cloud structure leakage from the UV images of clear-air density structure. The instrument requirements necessary to detect such structure and to discriminate aerosol-induced Mie scatter from Rayleigh scatter components consists of UV to visible band spectral imagers having sufficient spatial, temporal and spectral resolutions. Only moderate spectral resolution imagery in the 200 to 900 nm region over a range of sensor line of sight nadir angles is required to detect clouds and infer cloud types. However, high signal to noise ratios and high spatial resolution are required to characterize the structure power spectral density of clouds and clear-air scatter components. Middle atmosphere structure sounding capability on the mesoscale level allows connection between turbulent-like small scale atmospheric phenomenology and larger scale cloud-related and weather- driven atmospheric variability. We demonstrate the stratosphere-mesosphere sounding concept by applying a low altitude mesoscale stochastic structure (LAMSS) model. This model was derived from the NSS (non-stationary stochastic structure) model which utilizes multi-dimensional Fourier- space descriptions of wavelike, turbulent-like, and deterministic, large scale structure to simulate the effects of atmospheric earthlimb structure. LAMSS specifically address tropospheric background clutter processes such as clear-air wind shears, turbulence, temperature inversions, and cirrus cloud structure. The empirical models are applied to synthesis of visible, UV, and IR clutter backgrounds as measured by passive spectral imaging sensors such as the UVISI (UV, Visible Imagers and Spectral Imagers) sensors on the Mid-course Space Experiment (MSX). This paper analyzes images from MSX-UVISI to obtain cloud and atmospheric density structure characteristics in the 200 - 230 nm UV and 300 - 900 nm visible bands. These data illustrates the feasibility of the UV structure sounding concept by comparison to the synthesized structured backgrounds.